Cervical spondylotic myelopathy (CSM) has been a challenge for generations of surgeons. Although there have been no epidemiological studies on this disease, some reports indicate that the incidence of CSM could be as high as 23% [1] [2] among western populations and older patients [3] [4] , and would increase as these societies face an aging population [1] . In North America today, CSM is considered as the most common cause of spinal dysfunction in patients over the age of 55 [5] .

CSM is regarded as a progressively deteriorating process, due to cervical deformity and spondylotic changes with consecutive myelon compression [6] [7] . It is associated with gait disturbance such as spasticity or ataxia, paresthesia, hyperreflexia, radicular pain, weakness or stiffness of the legs, and sometimes neck pain [8] . The symptoms never follow a constant pattern and may be present all together or individually in CSM patients [9] .

Despite reports of successful conservative treatment of CSM, it is commonly considered as a surgical disease [6] . The first surgery for spinal decompression was performed by Sir Victor Horsley at the end of the 19^th Century [10] . Since then, many attempts to determine the optimal surgical approach have been made. Mummaneni et al. [11] recently reviewed the different surgical techniques for the treatment of CSM, such as dorsal decompression through laminectomy with or without fusion or laminoplasty, ventral decompression with or without fusion, or corpectomy. Which one might be the most favorable remains controversial though [7] . All of these techniques have the common goal of neuronal decompression while maintaining spinal stability in order to prevent further neurological deterioration [6] .

Although in recent years new surgical techniques have become popular, it has been postulated that laminectomy for dorsal decompression in a carefully selected group still needs to be considered as a choice of treatment of CSM [9] [12] [13] . The aim of this study is to evaluate the neurological outcomes and surgical complications of dorsal decompressive laminectomy without fusion in a selected consecutive series of patients presenting with CSM without preoperative clinical and radiological signs of instability.

Sixty-five consecutive patients, presenting with clinically confirmed CSM, who underwent dorsal decompression between 2000 and 2007 at the Department of Neurosurgery at the Kantonsspital Aarau were included in this single center study. The recorded clinical characteristics included gender, age, patient history, initial neurological status, surgical summaries, and clinical status at the time of follow-up examination. All patients showed cervical stenosis and compression of the myelon documented by MRI as a consequence of severe degenerative changes of the cervical spine. Cases with CSM secondary to trauma, neoplasm, or Chiari malformations were excluded. Data were analyzed and visualized with Microsoft Excel for Windows (Microsoft Corporation, Redmond, Wash- ington, USA) and GraphPad Prism version 4.00 for Windows (GraphPad Software, San Diego, California, USA), respectively. Values were expressed as mean ± SD.

The patients’ neurological status was obtained through physicians from our department at the time of admission and at a clinical follow-up examination. Neurological deficits were recorded for upper limbs (UL) and lower limbs (LL) as follows: gait disturbance, sensory loss, motor deficits, and hyperreflexia. Radiculopathy was assessed by patient history and physical examination. The postoperative clinical status was compared with the preoperative status as follows: worsened, unchanged, improved, and disappeared.

Radiological assessment included pre- and post-surgery magnetic resonance imaging (MRI), computed tomography (CT), and plain anterioposterior, lateral, and lateral flexion-extension radiographic views.

Standard laminectomy or laminoplasty was performed as described elsewhere [14] [15] . The decision for laminectomy or laminoplasty was made at the discretion of the surgeon on the basis of clinical and radiological findings. In all cases the decompression was completed under the microscope. No instrumentation or graft implantation was performed in this series, regardless of whether laminectomy or laminoplasty was performed. To minimize cervical spine movements, all patients underwent fiberoptic intubation. The perioperative use of a halo vest or stiffneck collar, additional cervical spine surgery at a later date, and postoperative complications were documented. All patients underwent follow-up examination at least once.

This study demonstrates that laminectomy without fusion for treatment of CSM resulted in overall improvement of neurological deficits in more than 70% of patients. In more than 80% of all cases, radicular pain was ameliorated. The risk of postoperative instability (1.5%) was low.

Frequencies of affected cervical levels: the number of patients concerned and their percentage of the whole group.

Frequencies of decompressed levels: the number of patients concerned and their percentage of the whole group.

A variety of surgical techniques for the treatment of CSM have been proposed over the years [11] . However, best practice guidelines have not yet been established. Recently, Mummaneni et al. reviewed data published between 1966 and 2007 to create an evidence-based approach considering the options and their efficacy in the surgical treatment of CSM [11] . Considering various techniques using both anterior and dorsal approaches, they established a series of recommendations. The authors concluded that laminectomy and anterior approaches such as anterior cervical decompression with fusion (ACDF) and anterior cervical corpectomy with fusion (ACCF) were equal surgical treatment options for CSM.

When deciding which treatment to use, the surgeon always needs to consider the different pathogeneses that can lead to CSM. Considering the pathophysiology of stenosis, it has been stated that patients with absolute cervical stenosis (<10 mm canal diameter) develop symptoms in their 40 s and 50 s [16] , while patients with relative stenosis (10 - 13 mm canal diameter) will present with radiculopathy and CSM in their 50 s and 60 s [7] . The patients’ age and co-morbidities and radiological assessments are important tools in planning the surgical approach. Hasegawa et al. [17] find that age in general does not seem to influence the surgical outcome of patients with CSM, but can influence the risk of perioperative complications. The pre-operative radiological assessment (which showed no signs of instability of the cervical spine) rather than consideration of age (mean 66 ± 13 years old) led to the decision for laminectomy without fusion in our series.

In recent years, anterior decompression has increased in popularity among surgeons despite higher rates of complications including dysphagia, hoarseness, or plate-related failures requiring reoperation [9] [18] [19] . The dorsal approach appears to be safe, but it has been linked with late onset of further clinical deterioration and spinal instability [20] [21] . In our series, only one patient developed instability of the cervical spine, whereas in a series of 58 patients undergoing multilevel laminectomy for CSM, reported by Guigui et al. [22] , 15% of the patients develop a destabilization of the spine postoperatively. In that report three patients underwent reoperation. In our study, 10% of our patients underwent additional ventral decompression, but this was part of a surgical plan designed to achieve adequate myelon decompression.

When surgery is performed on younger patients with additional ossification of the posterior longitudinal ligament, disc herniation, or spurs, ACDF or ACCF reduces further neurological deterioration and progression of spinal deformities. Dorsal decompression has been the surgical choice over many decades, especially in elderly patients suffering from CSM. Laminectomy, with or without fusion, and laminoplasty have been evaluated several times, and their efficacy, regarding CSM due to ossification of the yellow ligament, has been proven. It especially qualifies as a choice of treatment when addressing multilevel CSM [19] . Bapat et al. [23] conclude that for multilevel CSM the clinical outcome after anterior or posterior surgery is comparable. The complication rate is higher with anterior surgery. However, patients who undergo laminoplasty experience more continuous axial pain.

Due to the retrospective nature of our study, scores such as the Nurick scale [24] or the Japanese Orthopedic Scale were not applied and pain scores (Visual Analog Pain Scale) were not constantly assessed. However, individual pre- and post-operative clinical parameters were assessed in each case. The observed outcome was in line with reported results after dorsal decompression for CSM. Arnold et al. [25] report that of 44 patients treated with laminectomy, 77% show early improvement (within six months) and 52% show late improvement (mean: eight years). In a study with 129 patients, undergoing either anterior or posterior decompression, Bapat et al. [23] observe that 73% of the cases treated by laminectomy show an improvement. Reports from the 70s and 80s suggest a good outcome with recovery rates between 42% and 92% [15] [26] [27] . In terms of failure of surgery, defined as persistence or worsening of symptoms or neurological deficits, 9.2% of the patients undergo reoperation. This rate is similar to other studies reporting complication or failure rates from 6% to 38% [28] .

Based on the results of this study, laminectomy without fusion can be advocated as a safe and effective surgical strategy to treat CSM in patients without preoperative instability of the cervical spine. In these patients, post- procedural clinical deterioration and cervical spinal instability are reasonably low and overall improvement of neurological deficits and amelioration of radicular pain can be expected in a significant number of patients.